Comparisons of Clinical Outcomes in Patients Transferred to Combined Therapy with Peritoneal Dialysis and Hemodialysis over those Directly Transferred to Hemodialysis: A Prospective Multicenter Study in Japan

Background. Combined therapy with peritoneal dialysis (PD) and hemodialysis (HD) represents a treatment option for PD patients who cannot maintain adequacy of dialysis. Although several reports have indicated the clinical utility of combined therapy, most such studies have been small-scale, single-centered, retrospective and before-and-after test investigations lacking a control group. Methods. We conducted a prospective, multicenter, observational cohort study of 176 incident PD patients and compared patient survival and time-course changes in clinical parameters between patients switched from PD alone to combined therapy and patients switched directly to HD. Results. During a median follow-up of 41 months, 47 patients transferred to combined therapy and 35 patients transferred directly to HD. Among patients transferred to combined therapy, 66% transferred because of inadequate dialysis, and/or uid overload, compared to only 29% among patients directly transferred to HD. Five patients died after transfer to HD alone, whereas 1 patient died after transfer to combined therapy. Although mortality was greater among patients transferred to HD directly than among patients transferred to combined therapy, this difference disappeared after matching for uid overload and/or inadequate dialysis. dialysis


Background
Peritoneal dialysis (PD) is valuable for patients who have newly developed a need for renal replacement therapy (RRT) to preserve residual renal function (RRF), quality of life (QOL), and hemodynamic stability as compared to hemodialysis (HD) (1). However, the management of PD becomes more di cult as RRF declines and peritoneal function deteriorates, and inadequate dialysis and/or uid overload become clinically evident. Several guidelines recommend that optimal small-solute clearance, indicated as Kt/V ≥ 1.7, and euvolemia should be maintained by monitoring both urine volume and amount of ultra ltration achieved (2)(3)(4)(5).
Patients who become unmanageable by PD alone need to change the modality of RRT. Although such patients are often transferred directly to HD, transfer to combined therapy with PD and HD (generally consisting of 5-6 days of PD and one HD session per week) is another option. Combined therapy was introduced in the 1990s and has been rapidly spreading in Japan, and an estimated 1,800 patients (20% of all PD patients) have been on this therapy in 2017 (6). Combined therapy has been covered by the National Health Care Insurance System in Japan since April 2010. Several reports have demonstrated the clinical utility of combined therapy for correction of inadequate dialysis and/or uid overload (7-11). We established the Evaluation of the Adequacy of Renal replacement Therapy (EARTH) study group to clarify the clinical utility of combined therapy in 2008 and published a retrospective multicenter study of 104 patients switched from PD alone to combined therapy with PD and HD (7). However, previous studies, including that previous investigation, have targeted only patients starting combined therapy with PD and HD and have compared several clinical parameters between before and after switching to combined therapy.
The chief aim of this prospective, multicenter, observational cohort study including incident PD patients was to compare patient survival and several clinical parameters between patients switched from PD alone to combined therapy and those directly switched to HD.

Methods
We conducted a prospective, multicenter, observational cohort study that included 176 patients who started PD as rst RRT  Patient data, including age, sex, cause of end-stage renal disease (ESRD), comorbidities, physical examination, details of dialysis prescription and laboratory ndings at the start of PD and every year were registered using an anonymized case report form. The outcomes of interest were patient survival and changes in several physical and laboratory parameters. Regarding comorbidities, cardiovascular disease was de ned as heart failure, acute myocardial infarction, arrhythmia or conductive disturbance, valvular disease and other heart disease, whereas cerebrovascular disease was de ned as subarachnoid hemorrhage, cerebral hemorrhage, cerebral infarction and other cerebrovascular disease. Blood tests for hemoglobin, total protein, serum albumin, urea nitrogen, creatinine and β2 microglobulin were performed using standard laboratory techniques at each center. Since no de nite criteria for changing dialysis modality were applied, the timing of changing modality and selection of the type of the new modality, particularly combined therapy or HD alone, were left to the discretion of the treating physicians. If a patient changed dialysis modality, the date of transfer and reasons for changes (including inadequate dialysis, uid overload, both or other reasons) was required. Inadequate dialysis was de ned as laboratory test abnormalities including total Kt/V < 1.7, higher β2 microglobulin levels and appearance of uremic symptoms (including deterioration of nutritional status, decreased subjective global assessment (SGA) score, persistent anorexia, erythropoietin-stimulating agent (ESA)-hyporesponsive anemia and restless leg syndrome). Fluid overload was de ned as higher blood pressure (BP) or drug-resistant hypertension, heart enlargement or pleural effusion on chest X-rays and persistent peripheral edema or anasarca. The dialysate-to-plasma creatinine ratio (D/P Cr) was obtained from a peritoneal equilibration test (PET). In the PET as originally described, ultra ltration capacity is the net volume of ultra ltration achieved at 4 h using a 2.27% glucose exchange (3,5). Kt/V was calculated using guidelines from the Japanese Society for Dialysis Therapy (JSDT) (5).

Statistical analysis
Data are presented as mean ± standard deviation (SD) or median and interquartile range (IQR), as appropriate. Values of p < 0.05 were considered statistically signi cant. Differences among the three groups, including patients who continued PD, those who transferred to combined therapy with PD and HD and those who directly transferred to HD, were evaluated using one-way analysis of variance (ANOVA) or the non-parametric Kruskal-Wallis test, as appropriate. Chi-squared tests were used for the comparisons of nominal variables. Paired data were analyzed using the paired t-test or Wilcoxon signed-rank test, as appropriate. Comparisons of temporal changes in clinical parameters between patients transferred to combined therapy and those directly transferred to HD were assessed using repeated-measures two-way ANOVA. Kaplan-Meier survival analysis was used to compare patient survival between three groups. All statistical analyses were performed using JMP for Windows version 13.2.1 software (SAS Institute Inc., Cary, NC, USA). Table 1 lists baseline characteristics of 176 patients with incident PD (mean age, 60 ± 15 years; male, 71%). Causes of ESRD comprised chronic glomerulonephritis (CGN) in 56 patients (32%), diabetes in 66 (38%), hypertension in 27 (15%), polycystic kidney disease (PKD) in 5 (3%) and other or unknown in 22 (13%). Prevalences of cardiovascular disease, cerebrovascular disease and malignant disease were 12%, 9% and 6%, respectively. At the start of PD, urine volume, urea nitrogen and serum creatinine were 1230 mL/day (IQR 980-1800), 79.7 ± 28.6 mg/dL and 8.9 ± 2.4 mg/dL, respectively. Regarding PD-related tests, total, renal and PD Kt/V were 1.67 ± 0.49, 0.82 ± 0.41 and 0.84 ± 0.34, respectively. Baseline D/P Cr obtained from the PET was 0.60 ± 0.12.

Results
Median follow-up was 41 months (IQR, 25-67 months). Among the 176 enrolled patients, 82 patients were transferred from PD alone to other modalities, comprising 47 patients transferred to combined therapy with PD and HD and 35 patients directly transferred to HD, with times to transfer of 28 months (17-44 months) and 33 months (16-45 months), respectively. Among the 47 patients who transferred to combined therapy, 22 patients subsequently transferred to HD alone. Patients transferred to combined therapy were younger and had a higher body mass index (BMI) than those directly transferred to HD or those who continued PD alone, whereas no signi cant differences in other clinical parameters at the start of PD were seen between the three groups (Table 1). Reasons for changes in dialysis modalities are shown in Fig. 1. Among the patients transferred to combined therapy, 66% transferred because of inadequate dialysis, uid overload or both, whereas only 29% transferred because of these reasons among the patients directly transferred to HD (P < 0.01, Chi-squared test). Other reasons included inguinal hernia operation (n = 1), exit-site infection (n = 1), preparation for renal transplantation (n = 1), initiating dialysis with combined therapy (n = 1) and unknown reason (n = 12) among patients transferred to combined therapy, and di culty performing PD (n = 7), pleuroperitoneal communication (n = 4), peritonitis (n = 4), exit-site infection (n = 3), abdominal disorder or surgery (n = 3) and unknown reason (n = 4) among those directly transferred to HD. Regarding dialysis prescription, 93% of patients on combined therapy received 4-h HD using high-ux dialyzers once weekly, whereas the remaining 7% received 4-h online hemodia ltration once weekly. In general, PD was not carried out on the day of the HD session, and 35% of patients also did not perform PD on some other day each week, de ned as a 'PD holiday'. Table 2 shows comparisons of changes in clinical and biochemical parameters between patients transferred to combined therapy and those directly transferred to HD. BMI and urinary volume decreased after changing dialysis modalities in both groups. Urea nitrogen and serum creatinine decreased among patients directly transferred to HD, whereas no signi cant change was apparent among those transferred to combined therapy. BP decreased among patients transferred to combined therapy, but did not change signi cantly among those directly transferred to HD. Hemoglobin levels increased among patients transferred to combined therapy, whereas they no signi cant change was seen among those directly transferred to HD. Serum albumin and β2 microglobulin did not change signi cantly in either group. Overall, degrees of change in diastolic BP, urea nitrogen and serum creatinine were signi cantly different between groups (repeated-measured ANOVA P = 0.03, P < 0.01 and P < 0.01, respectively).
Thirty deaths occurred during follow-up (median, 41 months), including 11 cardiovascular deaths. One patient died after transfer to combined therapy, 5 patients died after transfer to HD alone and the remaining 24 patients died during PD therapy alone. Figure 2 shows a Kaplan-Meier analysis comparing patient survival between these three groups. (log-rank P < 0.01). Nineteen patients were censored due to renal transplantation and the median follow-up for that group was 13 months. Mortality was highest among patients continuing PD and was lower among those switching to combined therapy was among those switching directly to HD. Among patients switching dialysis modality because of inadequate dialysis and/or uid overload, 1 patient died after transfer to combined therapy and 1 patient died after transferring directly to HD alone. Kaplan-Meier analysis revealed no differences in patient survival between these two subgroups (log-rank P = 0.18, Fig. 3).

Discussion
We conducted a multicenter prospective observational cohort study to compare patient outcomes between patients who switched from PD alone to combined therapy with PD and HD and patients who switched directly to HD. No signi cant differences were observed with respect to survival outcomes between the two groups. We could not conduct more detailed analysis, e.g. multivariate analysis, to con rm the effects of dialysis modality on patient outcome due to the small number of outcomes, which was attributable to our inclusion of only patients who had newly started PD. To the best of our knowledge, this is the rst prospective case-control study in the eld of the combined therapy. Evidence has been accumulating that both inadequate dialysis and volume overload are signi cantly resolved after switching from PD alone to combined therapy (7-12). Our previous retrospective multicenter study included 104 patients who switched from PD alone to combined therapy with PD and HD, and revealed that body weight and serum creatinine decreased, whereas hemoglobin levels increased during 3 months of combined therapy (7). In addition, urinary volume decreased, whereas ultra ltration rate remained unchanged. Interestingly, D/P Cr decreased signi cantly, suggesting protective effects for the peritoneal membrane. However, all such studies, including our previous one, targeted only patients starting combined therapy with PD and HD, and thus lacked a control group.
We found that reasons for changing dialysis modality differed signi cantly between the two groups, with inadequate dialysis and/or uid overload as the reason in 66% of transfers to combined therapy, compared to only 29% for direct transfers to HD. These ndings revealed that patients suffering from inadequate dialysis or volume overload switched to combined therapy because they could continue PD. Conversely, the proportion of cases with this reason was lower and reasons for not continuing PD such as di culty performing PD, peritonitis and exit-site infection, pleuroperitoneal communication and abdominal disorder or surgery, were more frequent among patients who directly switched to HD. Interestingly, patients who directly transferred to HD were older and had a lower BMI than those who transferred to combined therapy at the start of PD. We therefore performed a subgroup analysis for only those patients changing dialysis modality due to inadequate dialysis and/or uid overload, i.e. those for whom PD was sustainable, and found that mortality among the two groups was similar. Further study with a larger number of outcomes is needed to compare clinical outcomes between patients who transfer to combined therapy and those who directly transfer to HD.
This study did not assess the e cacy of dialysis. The JSDT guideline (5) recommends that adequacy of dialysis should be determined using the concept of body uid clear space in combined therapy (5, 13), but established methods to assess the combine e cacy of PD and HD remain lacking. Although Kt/V was used for the assessment of dialysis adequacy both in PD and HD patients, its de nitions as applied these two treatments are totally different (14). We found that urea nitrogen and serum creatinine were decreased not among patients transferred to combined therapy, but among those directly transferred to HD. In addition, urinary volume after switching dialysis modality nearly disappeared. We thus speculated that dialysis adequacy for small-molecule uremic toxins was superior among patients who transferred directly to HD alone.
Several limitations to the present study must be considered. First, the number of outcomes was very small, as mentioned above. The comparison of patient outcomes, particularly using multivariate analysis, between patients who switched to combined therapy and those who switched directly to HD were therefore limited. Second, some degree of selection bias could not be avoided. because the timing of changing modality and selection of the type of new modality, especially in terms of combined therapy or HD alone, were left to the discretion of the treating physicians. On the other hand, a strength of this study was the multicenter design as a prospective observational cohort study of incident PD patients, allowing us to compare clinical outcomes between patients transferred to combined therapy and those transferred directly to HD.

Conclusions
This is the rst report comparing patient survival and changes in several clinical parameters among patients who transferred from PD alone to either combined therapy or directly to HD. Although the comparison of patient survival was di cult because of the relatively small number of deaths, PD patients suffering from inadequate dialysis and/or volume overload were clearly found to be able to continue PD therapy safely by switching to combined therapy. Further study with a larger number of outcomes is needed to con rm differences in clinical outcomes between these two groups.
Abbreviations PD, Peritoneal dialysis; RRT, renal replacement therapy; RRF, residual renal function; QOL, quality of life; HD, hemodialysis; EARTH, the Evaluation of the Adequacy of Renal replacement Therapy; ESRD, endstage renal disease; SGA, subjective global assessment; ESA, erythropoietin-stimulating agent, BP, blood pressure; D/P Cr, the dialysate-to-plasma creatinine ratio; PET, peritoneal equilibration test; JSDT, the Japanese Society for Dialysis Therapy; SD, standard deviation; IQR, interquartile range; ANOVA, analysis of variance; CGN, chronic glomerulonephritis; PKD, polycystic kidney disease; BMI, body mass index Declarations Ethics approval and consent to participate The ethics committee at Jikei University Hospital (approval number: 20-180 5470) and each participating hospital approved this study protocol, and all patients provided written informed consent to participate.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Funding This study was supported by grants from the Kidney Foundation, Japan. The funders played no roles in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.  Abbreviations: PD, peritoneal dialysis; HD, hemodialysis; BMI, body mass index; BP, blood pressure; CGN, chronic glomerulonephritis; PKD, polycystic kidney disease; Hb, hemoglobin; Alb, serum albumin; UN, blood urea nitrogen; Cr, creatinine; D/P Cr, dialysate-to-plasma ratio of creatinine. Abbreviations: HD, hemodialysis; BMI, body mass index; BP, blood pressure; Hb, hemoglobin; Alb, serum albumin; UN, blood urea nitrogen; Cr, creatinine. † Comparisons between before and after changing dialysis modality. ‡ Comparisons of the time course of changes in the values between patients transferred to combined therapy and those transferred to HD directly.